Mathematics Department, UCLA
T. Richthammer
winter 09, sheet 9
Feb 27, 2009
Homework assignments: Math 170A Probability, Sec. 1
109. Repeat problem 104, now assuming that the arrival time
X
of the bus is an exponential
RV with parameter
α
= 1
/
15.
Answer:
(a)
P
(
X >
10) =
R
∞
10
αe

αx
dx
=
e

10
/
15
≈
0
.
51.
(b)
P
(
X
≥
25

X >
15) =
P
(
X >
10)
≈
0
.
51. (memoryless!)
110. Repeat problem 105, now assuming the road has inﬁnite length, and
X
is exponentially
distributed with parameter
α
.
Answer:
E
(

X

a

) =
R
∞
0
α

x

a

e

αx
dx
=
R
a
0
(
a

x
)
αe

αx
dx
+
R
∞
a
(
x

a
)
αe

αx
dx
=
a
+
1
α
(
e

αa

1) +
1
α
e

αa
=
a
+
2
α
e

αa

1
α
=
f
(
a
), using
R
(
x

a
)
αe

αx
dx
=

(
x

a
)
e

αx

1
α
e

αx
(integration by parts). Minimizing
f
(
a
):
f
±
(
a
) = 0 gives
a
=
ln 2
α
, and we get
f
(
a
) =
ln 2
α
.
111.* In the lecture we have argued that under certain assumptions the number of events in
an interval [0
, t
] is a Poisson RV with parameter
αt
by using a discretization argument
and the Poisson approximation of binomial RVs. From this we have concluded that the
waiting time
Y
for the ﬁrst event is an exponential RV with parmeter
α
. We can also
argue directly, using the discretization and a corresponding approximation theorem on the
waiting times rather than on the number of successes: We divide the interval [0
,
∞
) into
pieces of equal length 1
/n
, and let
Y
n
denote the waiting time for the ﬁrst interval that
contains an event. Show the following:
(a) The assumptions from 6.3(d) imply that
Y
n
is a geometric RV with parameter
α/n
.
(b)
Y
n
/n
≈
Y
for large
n
.
(c) Whenever
Y
n
is a geometric RV with parameter
p
n
such that
np
n
→
α >
0 for
n
→ ∞
, the distribution function of
Y
n
/n
converges to the distribution function of
an exponential RV with parameter
α
at every point
c
.
(d)
Y
has to be an exponential RV with parameter
α
.
Answer:
(a) As in 6.3(d) we see that the RVs
X
i
(where
X
i
= 1 if an event occurs in the
i
th
interval, and
X
i
= 0 otherwise) form a Bernoulli sequence with parameter
p
=
α/n
.
Y
n
is
the waiting time for the ﬁrst success in the Bernoulli sequence and thus a geometric RV
with parameter
p
=
α/n
.
(b) If the ﬁrst success occurs in the
i
th interval (which has endpoints
i

1
n
and
i
n
), then
Y
n
=
i
and
Y
is in this interval, so

Y

Y
n
/n
 ≤
1
n
.
(c)
F
Y
n
/n
(
c
) =
P
(
Y
n
≤
cn
) =
P
(
Y
n
≤
c
n
), where
c
n
:=
±
cn
²
. We have
P
(
Y
n
≤
c
n
) =
∑
c
n
k
=1
p
n
(1

p
n
)
k
=
p
n
1

(1

p
n
)
cn
+1
1

(1

p
n
)
= 1

(1

p
n
)
c
n
+1
= 1

[(1

p
n
)
n
]
c
n
/n
, and in the
Poisson approximation theorem we have shown (1

p
n
)
n
→
e

α
, and we have
c
n
/n
=
±
cn
²
/n
→
c
. Thus
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview. Sign up
to
access the rest of the document.
 Winter '10
 RICHTHAMMER
 Normal Distribution, Binomial distribution, yn, normal RV, exponential RV

Click to edit the document details